Abstract

In this paper, we present a robust and mathematically sound ray-intersection algorithm for implicit surfaces. The algorithm is guar-anteed to numerically find the nearest intersection of the surface with a ray, and is guaranteed not to miss fine features of the surface. It does not require fine tuning or human choice of interactive parameters. Instead, it requires two upper bounds: "L" that limits the net rate of change of the implicit surface function f(x, y, z) and "G" that limits the rate of change of the gradient. We refer to an implicit surface with these rate limits as an "LG-implicit surface." Existing schemes to intersect a ray with an implicit surface have typically been guaranteed to work only for a limited set of implicit functions, such as quadric surfaces or polynomials, or else have been ad-hoc and have not been guaranteed to work. Our technique signifi-cantly extends the ability to intersect rays with implicit surfaces in a guaranteed fashion.

Citations